Palo Alto Networks NetSec-Architect Practice Test 18
Question 170
For an original educational Network Security Architect scenario, consider: A policy blocks an entire productivity suite because one embedded generative feature shares an existing host and is not correctly identified. Select TWO complementary steps: the targeted correction and its strongest prechange validation.
- Check: Compare actual application logs, cloud classification, affected tenant URLs, allow and block rules and user experience under a test policy.
- Check: Exercise representative inspection failures and regional failover while checking request routes, policy decisions, latency and alerting.
- Change: Include controlled AI red-team testing with representative misuse cases, prompt injection challenges, safety benchmarks and documented remediation before enabling privileged integrations.
- Change: Combine appropriate application identification and cloud-delivered classification features with granular AI application access policy and validation.
- Deny every HTTPS session to an entire vendor domain as the only way to control one AI feature inside the service.
Explanation
Correct Answer
A:Adaptive application identification can improve visibility of changing AI services, but decisions still require an explicit sanctioning and least-privilege policy. Acceptance: The targeted AI function is classified and handled according to business approval while unrelated authorized SaaS workflows continue.
D:Adaptive application identification can improve visibility of changing AI services, but decisions still require an explicit sanctioning and least-privilege policy. Evidence: Compare actual application logs, cloud classification, affected tenant URLs, allow and block rules and user experience under a test policy.
Incorrect Answers
B:Check: Exercise representative inspection failures and regional failover while checking request routes, policy decisions, latency and alerting. Adaptive application identification can improve visibility of changing AI services, but decisions still require an explicit sanctioning and least-privilege policy.
C:Change: Include controlled AI red-team testing with representative misuse cases, prompt injection challenges, safety benchmarks and documented remediation before enabling privileged integrations. Adaptive application identification can improve visibility of changing AI services, but decisions still require an explicit sanctioning and least-privilege policy.
E:Adaptive application identification can improve visibility of changing AI services, but decisions still require an explicit sanctioning and least-privilege policy.
Question 171
A network security architect is reviewing this environment: Developers paste unredacted customer identifiers into a sanctioned AI coding assistant despite a company policy forbidding such disclosure. A proposed shortcut is: Disable DLP for all enterprise AI applications because corporate authentication already proves the data is safe to upload. What should the security architect recommend as the narrowly scoped correction? Select ONE best answer.
- Apply enterprise DLP and AI-aware inspection for sensitive prompt and response content, aligning enforcement with sanctioned applications and documented exceptions.
- Disable DLP for all enterprise AI applications because corporate authentication already proves the data is safe to upload.
- Architect cluster-aware workload segmentation and suitable policy or runtime inspection for inference and agent components with narrowly defined communication paths.
- Select a Prisma AIRS deployment form factor and integration point consistent with who controls the model endpoint, network path and application API.
Explanation
Correct Answer
A:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
Incorrect Answers
B:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
C:This option addresses a different security architecture topic (Kubernetes AI workload segmentation) rather than the documented AI sensitive prompt DLP controls decision.
D:Select a Prisma AIRS deployment form factor and integration point consistent with who controls the model endpoint, network path and application API. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
Question 172
A Palo Alto Networks architecture review finds: Developers paste unredacted customer identifiers into a sanctioned AI coding assistant despite a company policy forbidding such disclosure. A design team proposes: Apply enterprise DLP and AI-aware inspection for sensitive prompt and response content, aligning enforcement with sanctioned applications and documented exceptions. Which architecture principle most strongly supports the recommended response? Select ONE best answer.
- An LLM agent becomes a privileged actor when it invokes tools; prompt content does not grant authorization to exceed identity, application or data entitlements.
- A security evaluation is version-specific; introducing new model weights, tools or data integrations can invalidate previous assurance assumptions.
- An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
- Compliance frameworks set obligations for data use and accountability; tool selection alone does not establish that an AI deployment complies with every jurisdiction.
Explanation
Correct Answer
C:Relevant validation before approval: Test synthetic sensitive records, matching classifiers, channel visibility, blocked requests, user notification and exception logging. Demonstrable success requires: Protected information is denied or masked according to classification while approved non-sensitive prompts complete normally.
Incorrect Answers
A:An LLM agent becomes a privileged actor when it invokes tools; prompt content does not grant authorization to exceed identity, application or data entitlements. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
B:A security evaluation is version-specific; introducing new model weights, tools or data integrations can invalidate previous assurance assumptions. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
D:Compliance frameworks set obligations for data use and accountability; tool selection alone does not establish that an AI deployment complies with every jurisdiction. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
Question 173
The architecture board has not yet authorized configuration changes. Current evidence indicates: Developers paste unredacted customer identifiers into a sanctioned AI coding assistant despite a company policy forbidding such disclosure. Which evidence best tests the architecture diagnosis before approving a disruptive change? Select ONE best answer.
- Match change tickets to artifact hashes, release configurations, risk assessments, red-team results and actual runtime policy coverage.
- Test synthetic sensitive records, matching classifiers, channel visibility, blocked requests, user notification and exception logging.
- Document personal-data classes, cross-border routes, access rights, logging and retention, control ownership and regulatory requirement mapping.
- Trace each model client, inference endpoint, control plane, identity boundary, payload path and failure behavior before selecting the integration method.
Explanation
Correct Answer
B:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel. Confirm afterward that: Protected information is denied or masked according to classification while approved non-sensitive prompts complete normally.
Incorrect Answers
A:Match change tickets to artifact hashes, release configurations, risk assessments, red-team results and actual runtime policy coverage. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
C:Document personal-data classes, cross-border routes, access rights, logging and retention, control ownership and regulatory requirement mapping. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
D:Trace each model client, inference endpoint, control plane, identity boundary, payload path and failure behavior before selecting the integration method. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
Question 174
The architecture team has implemented an approved correction and must now perform a production-safe acceptance test. Original problem: Developers paste unredacted customer identifiers into a sanctioned AI coding assistant despite a company policy forbidding such disclosure. The operations owner requires positive and negative test evidence and confirmation that legitimate business workflows still function. Which acceptance result demonstrates improved security without breaking authorized workflows? Select ONE best answer.
- The targeted AI function is classified and handled according to business approval while unrelated authorized SaaS workflows continue.
- Protected information is denied or masked according to classification while approved non-sensitive prompts complete normally.
- Untrusted text cannot cause unauthorized tool execution, and permitted tasks retain auditable identity and change approval.
- Risky action paths are blocked or constrained under approved test scenarios and the red-team findings are tracked to closure.
Explanation
Correct Answer
B:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
Incorrect Answers
A:The targeted AI function is classified and handled according to business approval while unrelated authorized SaaS workflows continue. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
C:Untrusted text cannot cause unauthorized tool execution, and permitted tasks retain auditable identity and change approval. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
D:Risky action paths are blocked or constrained under approved test scenarios and the red-team findings are tracked to closure. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
Question 175
A design sign-off request identifies the following problem: Developers paste unredacted customer identifiers into a sanctioned AI coding assistant despite a company policy forbidding such disclosure. Which paired security change and validation should close the design review? Select ONE best answer.
- Change: Apply enterprise DLP and AI-aware inspection for sensitive prompt and response content, aligning enforcement with sanctioned applications and documented exceptions. Verify: Test synthetic sensitive records, matching classifiers, channel visibility, blocked requests, user notification and exception logging.
- Change: Discover and classify AI application use, then apply role-aware sanctioned-app controls and documented data sharing restrictions. Verify: Review discovered AI applications, associated users, sanctioned status, DLP event samples and enforcement rules for enterprise versus personal accounts.
- Change: Include controlled AI red-team testing with representative misuse cases, prompt injection challenges, safety benchmarks and documented remediation before enabling privileged integrations. Verify: Examine the evaluation plan, attack prompts, guardrail failures, permission scope and retest results for the exact deployed agent release.
- Change: Architect AI runtime protection at a supported enforcement point that can inspect prompts and responses for policy-defined risks before sensitive content is released. Verify: Verify the actual application integration path, inspection coverage, data classification tests, blocking outcome and telemetry on permitted and rejected prompts.
Explanation
Correct Answer
A:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel. The change is successful only when: Protected information is denied or masked according to classification while approved non-sensitive prompts complete normally.
Incorrect Answers
B:Change: Discover and classify AI application use, then apply role-aware sanctioned-app controls and documented data sharing restrictions. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
C:Change: Include controlled AI red-team testing with representative misuse cases, prompt injection challenges, safety benchmarks and documented remediation before enabling privileged integrations. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
D:This option addresses a different security architecture topic (AI runtime inspection for prompt and response) rather than the documented AI sensitive prompt DLP controls decision.
Question 176
A network security architect is reviewing this environment: A legal department sends draft contracts to a model endpoint that lacks the required information handling agreement and needs classification-based blocking. A proposed shortcut is: Disable DLP for all enterprise AI applications because corporate authentication already proves the data is safe to upload. What should the security architect recommend as the narrowly scoped correction? Select ONE best answer.
- Combine appropriate application identification and cloud-delivered classification features with granular AI application access policy and validation.
- Disable DLP for all enterprise AI applications because corporate authentication already proves the data is safe to upload.
- Map AI model and application data flows to applicable regulatory obligations and risk controls, then choose architecture and enforcement based on documented requirements.
- Apply enterprise DLP and AI-aware inspection for sensitive prompt and response content, aligning enforcement with sanctioned applications and documented exceptions.
Explanation
Correct Answer
D:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
Incorrect Answers
A:Combine appropriate application identification and cloud-delivered classification features with granular AI application access policy and validation. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
B:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
C:Map AI model and application data flows to applicable regulatory obligations and risk controls, then choose architecture and enforcement based on documented requirements. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
Question 177
A Palo Alto Networks architecture review finds: A legal department sends draft contracts to a model endpoint that lacks the required information handling agreement and needs classification-based blocking. A design team proposes: Apply enterprise DLP and AI-aware inspection for sensitive prompt and response content, aligning enforcement with sanctioned applications and documented exceptions. Which architecture principle most strongly supports the recommended response? Select ONE best answer.
- An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
- AI-specific runtime threats can appear in model inputs, outputs and tool interactions; model scanning prior to release cannot inspect every later conversation.
- AI application security decisions depend on application identity, tenant and data handling policy, not on whether a URL contains the word artificial intelligence.
- A security evaluation is version-specific; introducing new model weights, tools or data integrations can invalidate previous assurance assumptions.
Explanation
Correct Answer
A:Relevant validation before approval: Test synthetic sensitive records, matching classifiers, channel visibility, blocked requests, user notification and exception logging.
Incorrect Answers
B:AI-specific runtime threats can appear in model inputs, outputs and tool interactions; model scanning prior to release cannot inspect every later conversation. This option addresses a different security architecture topic (AI runtime inspection for prompt and response) rather than the documented AI sensitive prompt DLP controls decision. The case-specific mechanism is: An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel. The relevant action is: Apply enterprise DLP and AI-aware inspection for sensitive prompt and response content, aligning enforcement with sanctioned applications and documented exceptions. Discriminating evidence: Test synthetic sensitive records, matching classifiers, channel visibility, blocked requests, user notification and exception logging.
C:AI application security decisions depend on application identity, tenant and data handling policy, not on whether a URL contains the word artificial intelligence. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
D:A security evaluation is version-specific; introducing new model weights, tools or data integrations can invalidate previous assurance assumptions. A legal department sends draft contracts to a model endpoint that lacks the required information handling agreement and needs classification-based blocking.
Question 178
The architecture board has not yet authorized configuration changes. Current evidence indicates: A legal department sends draft contracts to a model endpoint that lacks the required information handling agreement and needs classification-based blocking. Which evidence best tests the architecture diagnosis before approving a disruptive change? Select ONE best answer.
- Match change tickets to artifact hashes, release configurations, risk assessments, red-team results and actual runtime policy coverage.
- Test synthetic sensitive records, matching classifiers, channel visibility, blocked requests, user notification and exception logging.
- Inventory model versions, provenance, scan findings, ownership and deployment gates, then confirm that the scanned artifact matches the deployed release.
- Review discovered AI applications, associated users, sanctioned status, DLP event samples and enforcement rules for enterprise versus personal accounts.
Explanation
Correct Answer
B:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel. This check tests the actual control or routing boundary before a potentially disruptive change.
Incorrect Answers
A:Match change tickets to artifact hashes, release configurations, risk assessments, red-team results and actual runtime policy coverage. A legal department sends draft contracts to a model endpoint that lacks the required information handling agreement and needs classification-based blocking.
C:Inventory model versions, provenance, scan findings, ownership and deployment gates, then confirm that the scanned artifact matches the deployed release. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
D:Review discovered AI applications, associated users, sanctioned status, DLP event samples and enforcement rules for enterprise versus personal accounts. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
Question 179
The architecture team has implemented an approved correction and must now perform a production-safe acceptance test. Original problem: A legal department sends draft contracts to a model endpoint that lacks the required information handling agreement and needs classification-based blocking. The operations owner requires positive and negative test evidence and confirmation that legitimate business workflows still function. Which acceptance result demonstrates improved security without breaking authorized workflows? Select ONE best answer.
- Protected information is denied or masked according to classification while approved non-sensitive prompts complete normally.
- Recovery stays within the agreed business availability and risk requirements, with no hidden bypass of critical inspection on failover.
- Models with unresolved disallowed findings cannot be promoted, while the approved version retains documented scan evidence and exception ownership.
- Approved enterprise AI workflows remain usable and unsanctioned data transfers are logged or prevented per policy.
Explanation
Correct Answer
A:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel. The approved design step was: Apply enterprise DLP and AI-aware inspection for sensitive prompt and response content, aligning enforcement with sanctioned applications and documented exceptions.
Incorrect Answers
B:This option addresses a different security architecture topic (AI runtime resilience and fail behavior) rather than the documented AI sensitive prompt DLP controls decision.
C:Models with unresolved disallowed findings cannot be promoted, while the approved version retains documented scan evidence and exception ownership. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
D:Approved enterprise AI workflows remain usable and unsanctioned data transfers are logged or prevented per policy. An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.
Question 180
For an original educational Network Security Architect scenario, consider: A legal department sends draft contracts to a model endpoint that lacks the required information handling agreement and needs classification-based blocking. Select TWO complementary steps: the targeted correction and its strongest prechange validation.
- Change: Discover and classify AI application use, then apply role-aware sanctioned-app controls and documented data sharing restrictions.
- Check: Match change tickets to artifact hashes, release configurations, risk assessments, red-team results and actual runtime policy coverage.
- Check: Test synthetic sensitive records, matching classifiers, channel visibility, blocked requests, user notification and exception logging.
- Disable DLP for all enterprise AI applications because corporate authentication already proves the data is safe to upload.
- Change: Apply enterprise DLP and AI-aware inspection for sensitive prompt and response content, aligning enforcement with sanctioned applications and documented exceptions.
Explanation
Correct Answer
C:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel. Acceptance: Protected information is denied or masked according to classification while approved non-sensitive prompts complete normally.
E:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel. Evidence: Test synthetic sensitive records, matching classifiers, channel visibility, blocked requests, user notification and exception logging.
Incorrect Answers
A:Change: Discover and classify AI application use, then apply role-aware sanctioned-app controls and documented data sharing restrictions. A legal department sends draft contracts to a model endpoint that lacks the required information handling agreement and needs classification-based blocking.
B:Applied case: A legal department sends draft contracts to a model endpoint that lacks the required information handling agreement and needs classification-based blocking.
D:An allowed AI application does not imply all data is appropriate to submit; classification and policy must govern each disclosure channel.